重构相空间中孤立点的局部简单非线性降噪
发布时间：2018-09-12 17:17
【摘要】：为解决重构相空间中由于孤立点的存在而使简单非线性降噪法无法有效降噪的问题,本研究提出使用不同邻域半径实施多次局部简单非线性降噪的方法,去除重构相空间中的孤立点.使用该方法,对HZ公司的制造质量水平时间序列,在其重构相空间中经过实施16次邻域半径由大到小的孤立点去除操作后,在重构相空间中得到无孤立点的时间序列.经过本方法实施孤立点去除操作后的时间序列,只需再经过一次简单非线性降噪,就可以实现有效降噪,经过有效降噪的重构相空间才可以实施进一步的预测模型的拟合.
[Abstract]:In order to solve the problem that the simple nonlinear noise reduction method can not effectively reduce the noise due to the existence of isolated points in the reconstructed phase space, this paper proposes a method of multiple local simple nonlinear noise reduction using different neighborhood radii. Remove the isolated points in the reconstructed phase space. Using this method, the manufacturing quality level time series of HZ Company are obtained in the reconstructed phase space, after 16 times of outlier removal operations with large to small neighborhood radius, the time series without outliers is obtained in the reconstructed phase space. After the outliers are removed by this method, only one simple nonlinear de-noising is needed to realize the effective de-noising, and the fitting of the further prediction model can be carried out only after the reconstruction of the phase space of the effective de-noising.
【作者单位】： 华南理工大学工商管理学院;
【基金】：国家自然科学基金资助项目(51465020) 广东省软科学研究计划资助项目(2013B070206017) 广东省高等学校学科与专业建设专项资金科研类资助项目(2013WYXM0006)
【分类号】：O211.61
本文编号：2239690
[Abstract]:In order to solve the problem that the simple nonlinear noise reduction method can not effectively reduce the noise due to the existence of isolated points in the reconstructed phase space, this paper proposes a method of multiple local simple nonlinear noise reduction using different neighborhood radii. Remove the isolated points in the reconstructed phase space. Using this method, the manufacturing quality level time series of HZ Company are obtained in the reconstructed phase space, after 16 times of outlier removal operations with large to small neighborhood radius, the time series without outliers is obtained in the reconstructed phase space. After the outliers are removed by this method, only one simple nonlinear de-noising is needed to realize the effective de-noising, and the fitting of the further prediction model can be carried out only after the reconstruction of the phase space of the effective de-noising.
【作者单位】： 华南理工大学工商管理学院;
【基金】：国家自然科学基金资助项目(51465020) 广东省软科学研究计划资助项目(2013B070206017) 广东省高等学校学科与专业建设专项资金科研类资助项目(2013WYXM0006)
【分类号】：O211.61
【相似文献】
相关期刊论文 前1条
1 张雨;任成龙;;确定重构相空间维数的方法[J];国防科技大学学报;2005年06期
，本文编号：2239690
本文链接：https://www.wllwen.com/kejilunwen/yysx/2239690.html
